Accuracy of ultrasound-guided knee injections confirmed by fluoroscopy

Abstract

Background

Injection with viscosupplements is a common treatment for knee osteoarthritis. However, there is a lack of knowledge about how the injectate spreads within the knee following an injection.

Objective

Using ultrasound and fluoroscopy, this study seeks to assess whether injectate introduced into the suprapatellar recess disperses into the tibiofemoral joint.

Design

Descriptive case series and reliability test-retest study.

Setting

Outpatient rehabilitation center at an academic teaching hospital.

Patients

14 adults between 44 and 80 with knee osteoarthritis, defined as a grade 2–4 on the Kellgren and Lawrence scale, who were candidates for hyaluronic acid injections.

Interventions

Participants received ultrasound guided knee injections into the suprapatellar recess with hyaluronic acid and contrast. Post-injection fluoroscopic images were taken. The participants then underwent a walking protocol. Post-walking fluoroscopic images were then taken.

Main outcome measurements

Determining if an injectate introduced into suprapatellar recess localizes to the tibiofemoral joint following a walking test; and assessing interrater agreement with between 2 radiologists and 1 interventional physiatrist with regards to location of injectate.

Results

Injectate placed in the suprapatellar recess using ultrasound-guided technique will disperse to a varying extent from the suprapatellar recess into the tibiofemoral or patellofemoral joint after a brief bout of walking. Images of US-guided knee injections identified by an experienced interventionalist to represent correct needle placement and injectate location, confirmed by reference-standard fluoroscopy, can be corroborated by a blinded radiologist and are therefore reliable.

Conclusions

Fluoroscopic imaging confirmed that ultrasound-guided injection of hyaluronic acid into the suprapatellar recess dispersed into the tibiofemoral joint after a walking test. Future studies should examine whether the amount of injectate found in the tibiofemoral joint is correlated with patient outcomes.

Level of evidence

Level IV.

1. Introduction

Osteoarthritis is a leading cause of disability in the US and worldwide [1]. Intra-articular knee injections are a treatment option for patients with knee osteoarthritis who have not experienced symptomatic relief with more conservative measures such as physical therapy, orthotics, and oral pharmaceuticals and who are not yet appropriate surgical candidates. Common medications used for intra-articular injections include corticosteroids, hyaluronic acid, and orthobiologics. The ability of intra-articular injections to provide relief in patients with knee osteoarthritis is contingent upon the accurate delivery of the injectate into the joint space [2,3].

Intra-articular knee injections were initially performed using landmark-guided approaches before imaging confirmation was widely available. Techniques include the medial anterior and lateral anterior infrapatellar approach, which both enable direct access to the tibiofemoral synovial joint which is the main site of cartilage degradation in bicompartmental osteoarthritis. Other approaches include the superolateral patellar and superomedial patellar approaches in the extended knee. There are minimal soft tissue and few anatomical structures between the skin surface and the joint space using these approaches. However, these approaches can be problematic for several reasons: the accuracy and reliability of approach decreases with worsening osteoarthritis (decreased joint space), obesity (increased soft tissue) and joint anatomy variance. Additionally, these approaches carry a risk for damage to knee joint and related structures including menisci and ligaments, fat pad trauma, and risk for hemarthrosis with needle tip contacting periosteum and other vascular structures [4].

Ultrasound has become an increasingly popular and effective tool used to ensure accurate placement of an injectate. Extensive research has been performed to assess accuracy of landmark-guided knee injections, and the comparison of landmark-guided knee injections to ultrasound-guided knee injections. A case series performed by Lopes et al. of thirty-seven patients with rheumatoid arthritis is the only study to report 100% accuracy and efficacy of landmark-guided injections [5]. Since its publication, multiple studies have refuted accuracy of the landmark-guided knee injections. Systematic review of twenty three publications on landmark-guided viscosupplementation knee injection have varying accuracy depending on approach and experience of injector, with the superolateral patellar approach in the extended knee as most accurate at 87% accuracy [6,7]. Berkoff et al. completed a meta-analysis of thirteen studies, with results supporting ultrasound guided injections of the knee as more efficacious [8]. Sibbitt et al. performed two randomized control trials demonstrating ultrasound-guided injections had less procedural pain and better outcome, including longer lasting effect with more improvement [9,10].

Although ultrasound has helped clinicians in accurately identifying anatomical structures, the dispersion patterns of injectates, and the clinical significance of these dispersion patterns, still requires further investigation. For instance, past research has shown that in biceps tendon sheath injections, injectate did not flow into the glenohumeral joint despite their anatomical connection [11].

In the knee, the suprapatellar approach is commonly used for injections since the suprapatellar recess is easily visualized with ultrasound, and it is a commonly accepted principle that the suprapatellar recess communicates freely with the other parts of the knee intra-articular space [12]. However, it has never been shown using x-ray and radiocontrast dye confirmation that this communication actually occurs after medication is injected into the suprapatellar recess [13]. The purpose of this study was to determine if hyaluronic acid, a chondroprotective glycosaminoglycan molecule, disperses into the tibiofemoral joint following an ultrasound-guided injection,

2. Methods

2.1. Participants

Fourteen subjects volunteered to participate in this prospective study and were enrolled consecutively. The subjects ranged in age from 44 to 80 (sd ​= ​12.8yrs). Eight of the subjects were female, and six of the subjects were male. Patients were enrolled from April 16, 2015 to June 14, 2016. All participants provided written informed consent after completion of a consent process approved by the institutional review board. The subjects had to meet the American College of Rheumatology classification criteria for osteoarthritis of the knee, which includes radiographic evidence for osteophytes and at least one of the following three items: age greater than or equal to 50 years old, morning stiffness less than or equal to 30 ​min in duration, and crepitus on motion. The subjects enrolled in this study were all categorized as between grade 2–4 on the Kellgren and Lawrence Osteoarthritis grading scale [14]. Exclusion criteria included the following groups of patients: those with a reason to avoid exposure to radiation (e.g. pregnancy), those who otherwise meet exclusion criteria for undergoing fluoroscopy (e.g. weight >400 lbs, inability to remain still, inability to transfer onto and off a narrow fluoroscopy table), those with active infections or skin diseases in the area of the injection site, those with allergy or another contraindication to contrast dye or to hyaluronic acid, those unable to give consent (e.g. cognitive impairment), and non-English speaking subjects.

2.1.1. Intervention

All participants had pre-injection fluoroscopic images taken, and they then received ultrasound-guided injections performed by the principal investigator. Injections were performed using an in-plane sagittal oblique approach towards the suprapatellar recess. Lidocaine was initially injected, and after a bursal spread was identified, the syringe was exchanged for a solution of hyaluronic acid that has been premixed with contrast. Connector tubing was used to avoid the inadvertent movement of the needle or needle tip out away from the intended target. Immediate post-injection fluoroscopy images were then taken. The patients then walked for 3 ​min, and did ten sit-to-stand maneuvers from a chair. After this, post-walking fluoroscopy images were taken. The anatomical location of the injectate was then assessed by two radiologists and one interventional physiatrist.

2.1.2. Statistical analysis

Participant baseline characteristics were described using means and standard deviations. To determine the diagnostic accuracy of the injections and inter-rater agreement, data was analyzed using the IBM SPSS statistical software (Chicago, IL). A cross-tabulation kappa analysis was used to determine the inter-rater agreement, and a cross-tabulation analysis was used to determine the diagnostic accuracy of the injections.

3. Results

Thirty-two potentially eligible subjects were assessed for participation in the study. Of these potentially eligible subjects, eight were excluded. Two were excluded due to allergies, and six declined to participate in the study. The average age of participants was 66 years old (sd 12.82) and 57% of the participants were female (Table 1). The average KL-OA Grade was 2.7 (sd 0.82) (Table 1).

Table 1.

Demographics and KL-OA grading of participants.

Participant Number	Age (years)	Gender	KL-OA Grade
Individual 1	74	F	2
Individual 2	48	F	2
Individual 3	79	M	2
Individual 4	72	M	2
Individual 5	79	M	3
Individual 6	49	M	4
Individual 7	80	F	2
Individual 8	55	F	2
Individual 9	74	F	4
Individual 10	63	M	4
Individual 11	76	F	3
Individual 12	58	F	3
Individual 13	73	F	3
Individual 14	44	M	2
Average age:	66	St dev:	12.82425586
Female %	57.1428571
Average grading:	2.71428571	St dev:	0.825420306

The fourteen participants underwent the ultrasound guided knee injection and had their ultrasound and fluoroscopic images interpreted by two radiologists and one interventional physiatrist. Each of the images were scored based on accuracy of the needle insertion or contrast die. A score of zero indicated it was unclear if in the desired location, a score of 1 if not in the desired location, and a score of 2 if in the desired location. A composite score was then created (Table 2).

Table 2.

Scoring for ultrasound guided knee placement, pre-walk, and post walk, performed by two radiologists and an interventional physiatrist.

Participant	Ultrasound, Pre-Walk, or Post-Walk	Scoring
		Radiologist 1	Radiologist 2	Interventional Physiatrist	Composite
Individual 1	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 2	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 3	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 4	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	0	0	2
Individual 5	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 6	Ultrasound	0	2	2	4
	Pre-Walk	1	2	2	4
	Post-Walk	2	2	2	6
Individual 7	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 8	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 9	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 10	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 11	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	0	4
Individual 12	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 13	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6
Individual 14	Ultrasound	2	2	2	6
	Pre-Walk	2	2	2	6
	Post-Walk	2	2	2	6

For thirteen of the fourteen participants, all three image reviewers agreed that the needle placement had been accurately placed in the suprapatellar recess on the initial ultrasound image (Fig. 1). On the pre-walk fluoroscopic imaging, all three image reviewers agreed that contrast was visible in the suprapatellar recess in all fourteen of the cases (Fig. 2). The post-walk fluoroscopic imaging results were as follows: one of the reviewers visualized contrast in the tibiofemoral or patellofemoral joint spaces in all fourteen cases; another reviewer visualized contrast in the tibiofemoral or patellofemoral joint spaces in thirteen of fourteen cases; and the last reviewer saw the contrast in the tibiofemoral or patellofemoral joint spaces in twelve of the fourteen cases (Fig. 3).

Fig. 1.

Needle placement in the suprapatellar recess under ultrasound guidance. A ​= ​patella; B ​= ​suprapatellar recess; C ​= ​needle entry.

Fig. 2.

Pre-ambulation contrast spread. A ​= ​Distal Femur, B= Patella, C = Proximal Tibia.

Fig. 3.

Post-ambulation contrast spread. A ​= ​Distal Femur, B= Patella, C = Proximal Tibia.

Statistical interpretation of these results demonstrated that injectate placed in the suprapatellar recess using ultrasound-guided technique will disperse from the suprapatellar recess into the tibiofemoral or patellofemoral joint after a brief walk with sensitivity of 96.3% (sd 2.07%, confidence interval 90.8%–99.1%) and specificity of 50% (sd 35.36%, confidence interval 3.8%–96.2%). Furthermore, these results show that images of ultrasound-guided knee injections identified by an experienced interventionalist to represent correct needle placement and injectate location, confirmed by reference-standard fluoroscopy, can be corroborated by a blinded radiologist and are therefore reliable (93%–100% agreement, kappa 0.632, p ​= ​0.011).

4. Discussion

This prospective case series of fourteen patients demonstrates hyaluronic acid introduced into the suprapatellar recess disperses into the tibiofemoral and patellofemoral joint by ultrasound-guided injection, which has not been previously demonstrated. The use of fluoroscopy to confirm the dispersion of hyaluronic acid from the suprapatellar recess to the joint space validates ultrasound as an effective modality and the suprapatellar recess as an ideal anatomical location for intra-articular knee injections in patients with osteoarthritis [6], A study performed by Zuber et al. supported the superolateral approach as providing the most direct access to the synovial joint [15]. Ultrasound-guided injections also offer a clinically significant alternative to landmark-guided injections because they provide a more benign adverse effect profile. Specifically, they carry a decreased risk of meniscal damage, ligamentous injury and intravascular injection. One prior study demonstrated patients receiving ultrasound-guided knee hyaluronic acid injections were significantly less likely to undergo subsequent knee arthroplasty than those who underwent a landmark-guided knee injection [16]. Furthermore, the accuracy of these injection is affected less by increased soft tissue, which is seen with obesity, and narrowing of the joint space, which is seen in advanced osteoarthritis.

Although contrast was seen in the tibiofemoral space, one could argue it appears small and unconcentrated on the images provided (Fig. 2, Fig. 3). However, this study supports targeting the suprapatellar recess can achieve medicine spreading into the tibiofemoral space. There is no study demonstrating medication in the tibiofemoral space stays concentrated there and does not push cephalad towards the suprapatellar recess after injection.

The lack of a control group who received the hyaluronic acid by means of a palpation guided injection is a limitation in this study. Future research should seek to compare the localization of an injectate in patients who receive ultrasound guided and palpation guided injections. There is also possibility dispersion may have occurred without brief walking, which can be addressed in future studies by assessing the degree of injectate spread when individuals ambulate after injection versus do not ambulate. The findings of such a study could guide clinical practice, such as recommending patients ambulate post-injection to increase dispersion into the joint space. Another limitation of this study was that the amount of injectate in the joint space was not quantified. Future projects should attempt to quantify the amount of injectate in the joint space and see if this is correlated with symptomatic improvement. Additionally, an effort should be made to see if the severity osteoarthritis affects the dispersion of injectate. Future analyses should also explore the effect of injectate viscosity of dispersion patterns.

5. Conclusion

This prospective case series demonstrates injectate introduced into the suprapatellar recess disperses into the tibiofemoral and patellofemoral joint. The accuracy of this finding was shown using ultrasound and fluoroscopic imaging, and the reliability of this data was shown by demonstrating the inter-rater agreement between two radiologists and one interventional physiatrist. Other characteristics that may affect dispersion, like osteoarthritis severity, injectate viscosity, and the location of the injection should be studied in the future.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.